fix(linux): enable PulseAudio/PipeWire system audio capture

[thalesac]

Summary

• Fix Linux system audio capture by properly enumerating PulseAudio/PipeWire monitor sources
• Previously, only microphone audio was captured on Linux because cpal's ALSA backend doesn't see PulseAudio monitor sources

Changes

• linux.rs: Enumerate monitor sources via pactl list sources short command instead of relying on cpal's ALSA backend
• configuration.rs: Use PULSE_SOURCE env var with 'pulse' ALSA device for capturing from PulseAudio sources
• speakers.rs: Add default_system_audio_device() for Linux that finds the best available monitor source
• recording_manager.rs: Use Linux-specific system audio device selection

How it works

1. Query pactl for .monitor sources (system audio capture points)
2. Set PULSE_SOURCE environment variable to the selected monitor
3. Use the pulse ALSA device which routes through PulseAudio
4. Fall back to ALSA loopback devices if pactl unavailable

Test plan

• Build with pnpm run tauri:build on Linux
• Verify PulseAudio monitors appear in Settings > Audio Devices > System Audio dropdown
• Select a monitor source (e.g., "Built-in Audio" / alsa_output.pci-*.monitor)
• Start recording while playing YouTube/audio
• Verify both microphone AND system audio are captured

System tested

• Ubuntu/Linux with PipeWire (PulseAudio compatibility layer)
• NVIDIA GPU with CUDA acceleration

🤖 Generated with Claude Code

[thalesac]

disclaimer: vibe coded - fyi @sujithatzackriya